[R-sig-Geo] About fixed and adaptive kernel of GWmodel version 1.2-5

[Vinay Elothunkal]

Hi all,

I am using "gwr.predict" to estimate coefficients from multispectral bands
using rgrass7. Adaptive=T provides me better accuracy results that fixed
bandwidth for randomly distributed Y variables.

   In my second case,  Y variables are not randomly distributed but its
like an arbitrary track. While evaluating the result, adaptive bandwidth
provides demarcation along the track like[1]  and fixed bandwidth result [2
]provides no such demarcation and produce  better estimation in terms R2.
please note that  the black track shown on the [1] and [2] are the Y
variables used for "gwr.predict".


   Can anyone tell why this kind of results providing by the adaptive
kernel, is this obvious or abnormal?
[1] Adaptive_BW1.jpg
[2] Fixed_BW2.jpg


Any kind of comments related GWmodel is appreciated

Regards,
Vinay
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20160315/a4563f11/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Adaptive_BW1.jpg
Type: image/jpeg
Size: 13233 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20160315/a4563f11/attachment.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Fixed_BW2.jpg
Type: image/jpeg
Size: 14087 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20160315/a4563f11/attachment-0001.jpg>